<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#6600CC">
Dear Damir, Colleagues<br>
<br>
if I may put in my two cents (or my two euros!!) in this discussion<br>
1) I assume that we have to interpret the previous message from a US
perspective; As you are in the US , the fair use is a very common
doctrine, unfortunately you can not use such arguments in a large
number of countries in particular in Europe; UK has moved recently
to this but not clear yet.<br>
2) in fair use , the first of the four factors is <b>(1) the
purpose and character of the use, including whether such use is of
a commercial nature or is for nonprofit educational purposes</b>;
so for research and educational purposes (including sharing data for
research) I second Orion arguments.<br>
<br>
3) the common misinterpretation is "can you trace the sources of my
"list of words" or what ever "derived outcomes" from the copyrighted
data; my legal advisors insist often that in "Copyright
infringement" the key word is "Copy" , so as long as you copy data
(that is what we do when we crawl/harvest etc. you are already
infringing the Copyright law.<br>
<br>
But the other key word that my lawyers utter more often is "risk",
what is the risk (i.e. what would be the benefit for copyright
owners to sue you!!), I guess you are the only one who can assess
this ; in Europe everyone is trying to sue Google , Microsoft etc.
!!<br>
<br>
all the best<br>
Khalid<br>
<br>
<br>
<br>
<div class="moz-cite-prefix">On 2015-01-06 07:15, Orion Montoya
wrote:<br>
</div>
<blockquote
cite="mid:CAE7YMx8t5o0UR8H6rgyHJwTFYVvc_PUumQXY-UiH2bBypiDUgA@mail.gmail.com"
type="cite">
<div dir="ltr">Word lists and frequency profiles would seem to be
safely in the realm of fair use: <a moz-do-not-send="true"
href="http://en.wikipedia.org/wiki/Fair_use" target="_blank">http://en.wikipedia.org/wiki/Fair_use</a>
. The Google Books Ngrams data, distributed up to 12-grams by
Google, are one example of people distributing rather high-N
ngrams. Of course Google fought with the Authors Guild over
Google Books in general, but I don't recall this distribution of
ngram data being part of their fight, and in the end the Authors
Guild didn't win the obscurity they were pleading for. For
another example, <a moz-do-not-send="true"
href="http://commoncrawl.org/" target="_blank">http://commoncrawl.org/</a>
distributes a massive crawl of the web for researchers (or
anybody) which is far more wholesale copying+redistribution than
you're proposing, but they follow the normal rules that
webcrawlers follow and are doing just fine (and are a very
useful resource!).
<div><br>
</div>
<div>So I would personally have zero legal worry about what
you're proposing. I would have no qualms about either academic
research or commercial applications (or commercial
distribution) of that derived data. Adam is (as usual) right,
that you shouldn't even ask anybody for permission.
<div><br>
</div>
<div>The thing about fair use that can make university lawyers
uncomfortable is that it's an "affirmative defense" -- you
can argue it in court if someone sues you, but there's no
guarantee that you can use it to stay out of court in the
first place, which can be expensive.</div>
<div><br>
</div>
<div>But the other thing about the fair use defense is that,
in order for you to use it, somebody needs to be able to
claim that you're infringing their copyright in the first
place. If you're just distributing frequency lists, there's
no trace of a copyrighted work to be found; even at the
5-gram level, it's very hard to find any actionable
infringement: the fourth principle to be considered in
evaluating fair use is "<span
style="color:rgb(37,37,37);font-family:sans-serif;font-size:13px;line-height:21.2800006866455px;background-color:rgb(249,249,249)">the
effect of the use upon the potential market for or value
of the copyrighted work" and in your case that effect
should be exactly nil.</span></div>
<div><br>
</div>
</div>
<div>You could save yourself a bit of busywork, and maybe offer
your university's lawyers some psychological insulation from
legal risk, by using existing corpora resources like Common
Crawl.</div>
<div><br>
</div>
<div>Part C of your question --- "are there jurisdictions where
this might be illegal" --- is the fuzziest to answer; the
Berne Convention allows signatory countries to define fair use
for themselves, so there might be jurisdictions where this
could be risky, but they're probably places for which it's
challenging to get a visa anyway. I am not a lawyer, just a
copyright geek and a subscriber of "5 Useful Articles" by
Parker Higgins & Sarah Jeong, <a moz-do-not-send="true"
href="http://tinyletter.com/5ua">http://tinyletter.com/5ua</a>
, an amusing and edifying weekly email about the inherent
comedy of US IP law in the 21st century.</div>
<div><br>
</div>
<div>Cheers,</div>
<div><br>
</div>
<div>Orion</div>
<div class="gmail_extra"><br>
<div class="gmail_quote">On Mon, Jan 5, 2015 at 9:30 PM, Adam
Kilgarriff <span dir="ltr"><<a moz-do-not-send="true"
href="mailto:adam.kilgarriff@sketchengine.co.uk"
target="_blank">adam.kilgarriff@sketchengine.co.uk</a>></span>
wrote:<br>
<blockquote class="gmail_quote" style="margin:0px 0px 0px
0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex">
<div dir="ltr">Dear Damir,
<div><br>
</div>
<div>a few thoughts:
<div><br>
</div>
<div>In an innocent world view, the law says what is
allowed and what is not. The more I see of how the
legal profession works, the clearer it is that it's
all political, in the sense that the judgements that
build the case law (at least in UK) are made based
on how well the lawyers played the game, how much
money was involved, who had a sniff of how much
money they might make.</div>
<div><br>
</div>
<div>It's not about what is legal (which is always, in
this area, underspecified), it is about risk
management.</div>
<div><br>
</div>
<div>If no-one sees a money-making opportunity, there
is very little legal risk since no-one will take you
to court. </div>
<div>If you're a big organisation, you can always be
taken to court and sued for large sums. This has
had horrible consequences for the JISC group at
ISPRA: they are part of the EU, a very large
organisation, and have had their work restricted by
ambulance-chasing lawyers with a glint in their eyes
for winning plump settlements. </div>
<div><br>
</div>
<div>What you might be willing to do personally -
given that you are probably, not, as an individual,
worth suing, and your motivation for doing
interesting work is high - is very different to what
a (probably) rich organisation like your university
might be willing to do. If you want to do
something, don't ask! (Specially not the university
lawyers. You'll probably never get an answer - even
more frustrating than a simple 'no'.)</div>
</div>
<div><br>
</div>
<div>Sorry if that is not very helpful</div>
<div><br>
</div>
<div>Adam</div>
<div><br>
</div>
</div>
<div class="gmail_extra"><br>
<div class="gmail_quote">
<div>
<div>On 6 January 2015 at 04:00, Damir Cavar <span
dir="ltr"><<a moz-do-not-send="true"
href="mailto:dcavar@me.com" target="_blank">dcavar@me.com</a>></span>
wrote:<br>
</div>
</div>
<blockquote class="gmail_quote" style="margin:0px 0px
0px
0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex">
<div>
<div>Hi everybody,<br>
<br>
I know, this question has been addressed a lot,
but, just to get an<br>
update on this issue and your expert opinion:<br>
<br>
If I am accessing the internet from the US, as I
am right now, and I<br>
decide to generate N-gram-based language models
by exploiting the web as<br>
a corpus and publish the word-lists and
frequency profiles openly on my<br>
homepage, sell them even, change or manipulate
them, and reuse them in<br>
various ways, would this be<br>
<br>
a. ok as fair-use for research only, excluding
commercial use<br>
b. legal in general, independent of my research
interests<br>
c. legal only in some countries (so, my models
would be illegal in some<br>
others)<br>
<br>
What is the current status of the web as a
corpus and extracted language<br>
models from the legal perspective in the US and
globally?<br>
<br>
If I do the same now with open-access journals
and extract frequency<br>
profiles of tokens for a certain research
domain, would it be the same?<br>
It I use Google Books? Or even some news
website?<br>
<br>
Is the extraction of a language model, maybe a
domain specific frequency<br>
profile a copyright infringement per se? The
text cannot be<br>
reconstructed, the content is not visible, the
authors style neither, in<br>
particular not, if the corpus is larger etc.<br>
<br>
Thanks!<br>
<br>
Damir<br>
<br>
<br>
<br>
--<br>
Damir Cavar<br>
Department of Linguistics<br>
Indiana University<br>
<br>
<br>
<br>
</div>
</div>
_______________________________________________<br>
UNSUBSCRIBE from this page: <a
moz-do-not-send="true"
href="http://mailman.uib.no/options/corpora"
target="_blank">http://mailman.uib.no/options/corpora</a><br>
Corpora mailing list<br>
<a moz-do-not-send="true"
href="mailto:Corpora@uib.no" target="_blank">Corpora@uib.no</a><br>
<a moz-do-not-send="true"
href="http://mailman.uib.no/listinfo/corpora"
target="_blank">http://mailman.uib.no/listinfo/corpora</a><br>
<br>
</blockquote>
</div>
<span><font color="#888888"><br>
<br clear="all">
<div><br>
</div>
-- <br>
<div>
<div dir="ltr">=============================================<br>
<a moz-do-not-send="true"
href="http://www.kilgarriff.co.uk/"
target="_blank">Adam Kilgarriff</a>
<a moz-do-not-send="true"
href="mailto:adam@sketchengine.co.uk"
target="_blank">adam@sketchengine.co.uk</a>
<br>
Director <a
moz-do-not-send="true"
href="http://www.sketchengine.co.uk/"
target="_blank">Lexical Computing Ltd</a>
<br>
Visiting Research Fellow <a
moz-do-not-send="true"
href="http://leeds.ac.uk/" target="_blank">University
of Leeds</a>
<div><i><font color="#006600">Corpora for all</font></i> with <a
moz-do-not-send="true"
href="http://www.sketchengine.co.uk/"
target="_blank">the Sketch Engine</a> and
<a moz-do-not-send="true"
href="http://skell.sketchengine.co.uk/"
target="_blank">SKELL</a> <i>
</i></div>
<div>=============================================</div>
</div>
</div>
</font></span></div>
<br>
_______________________________________________<br>
UNSUBSCRIBE from this page: <a moz-do-not-send="true"
href="http://mailman.uib.no/options/corpora"
target="_blank">http://mailman.uib.no/options/corpora</a><br>
Corpora mailing list<br>
<a moz-do-not-send="true" href="mailto:Corpora@uib.no"
target="_blank">Corpora@uib.no</a><br>
<a moz-do-not-send="true"
href="http://mailman.uib.no/listinfo/corpora"
target="_blank">http://mailman.uib.no/listinfo/corpora</a><br>
<br>
</blockquote>
</div>
<br>
</div>
</div>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset>
<br>
<pre wrap="">_______________________________________________
UNSUBSCRIBE from this page: <a class="moz-txt-link-freetext" href="http://mailman.uib.no/options/corpora">http://mailman.uib.no/options/corpora</a>
Corpora mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Corpora@uib.no">Corpora@uib.no</a>
<a class="moz-txt-link-freetext" href="http://mailman.uib.no/listinfo/corpora">http://mailman.uib.no/listinfo/corpora</a>
</pre>
</blockquote>
<br>
<div class="moz-signature">-- <br>
<p> ************************************************* <br>
<b> Khalid CHOUKRI </b> <br>
ELRA General secretary & ELDA CEO <br>
email: <a class="moz-txt-link-abbreviated" href="mailto:choukri@elda.org">choukri@elda.org</a> ; Web: <a class="moz-txt-link-abbreviated" href="http://www.elra.info">www.elra.info</a> <a class="moz-txt-link-abbreviated" href="http://www.elda.org">www.elda.org</a> <br>
Tel. +33 1 43 13 33 33 - Fax. +33 1 43 13 33 30 <br>
*************************************************** <br>
** <b> Info on LREC: <a class="moz-txt-link-abbreviated" href="http://www.lrec-conf.org">www.lrec-conf.org</a> <b><br>
**************************************************** </b></b></p>
<b><b>
</b></b></div>
</body>
</html>